Objective: To evaluate and compare the predictive powers of clinical examination, electroencephalography (EEG), and studies of short-latency somatosensory evoked potentials in determining the prognosis in anoxic coma.
Design: Prospective case series of patients in anoxic coma, whose prognoses were uncertain based on previously established clinical criteria. The clinical features, EEG, and somatosensory evoked potentials results were correlated with outcome.
Setting: A 40-bed intensive care unit in a university teaching hospital.
Patients: Thirty-four consecutive patients admitted over a 2-yr period with anoxic coma as the principal diagnosis.
Interventions: None.
Measurements and main results: Twenty-seven (79%) patients never recovered consciousness and seven (21%) patients made a good recovery. One of six patients whose pupillary reflexes were present but whose other cranial nerve reflexes were absent on day 1 recovered, but none of the seven patients with these features still present on day 3 survived. None of the patients with motor responses of extension to painful stimuli or worse on days 1 or 3 recovered. The EEGs were classified into benign, uncertain, and malignant categories. The results of both EEG and somatosensory evoked potentials studies were strongly associated with outcome. With malignant EEG, the sensitivity was 74%, the specificity was 71%, and the positive predictive value was 90% [corrected] for the prediction of no recovery (death or persistent vegetative state). However, two patients with an initially malignant EEG eventually made a good recovery. The sensitivity for low amplitude or absent somatosensory evoked potentials for prediction of no recovery was 66%. There were no falsely pessimistic predictions with somatosensory evoked potentials, as all 18 patients with absent or low-amplitude responses had no recovery (specificity and positive predictive value were 100%). EEG and somatosensory evoked potentials studies were complementary to clinical examination in the determination of prognosis. Using a combined clinical and electrophysiologic approach, 63% of patients who had no recovery could be identified by day 3. Repeat EEG and somatosensory evoked potentials studies were of value in patients whose prognoses were uncertain, as their evolution invariably correlated with outcome.
Conclusions: Based on the present data and a literature review, we propose that clinical examination combined with the results of EEG and somatosensory evoked potentials can be used to establish an early, definitive prognosis in a significant proportion of patients in anoxic coma. On day 3 or thereafter, patients with motor response of extension to pain or worse and malignant EEG, or those patients with flexor posturing or worse and bilaterally absent cortical somatosensory evoked potentials invariably have poor outcome. However, some patients with initially malignant EEG and normal somatosensory evoked potentials may recover and should be supported until their prognoses become more definitive.